.NET 7 - Process terminated. Error while reaping child. errno = 10

[JaroslavMajera]

Description

Hello. We have a cloud docker microservice which runs headless chromium in process. When we switch to .NET 7 we have encountered Process terminated. Error while reaping child. errno = 10 during killing chromium process. We cannot reproduce this behavior with .NET 6. I have prepared repro https://github.com/JaroslavMajera/DotNetReaper.

Thanks

Reproduction Steps

1. Prepare Dockerfile which periodically runs chromium or maybe something else which spawns child processes.
2. Start chromium process
3. Call Process.Kill a Process.WaitForExit()
4. Repeat

Source code: DotNetReaper-main.zip

Expected behavior

No Exception and chromium is successfully killed

Actual behavior

Sometimes we get

Process terminated. Error while reaping child. errno = 10 at System.Environment.FailFast(System.String) at System.Diagnostics.ProcessWaitState.TryReapChild(Boolean) at System.Diagnostics.ProcessWaitState.CheckChildren(Boolean, Boolean) at System.Diagnostics.Process.OnSigChild(Int32, Int32)

Regression?

Cannot reproduce on .NET 6

Known Workarounds

No response

Configuration

• .NET 7
• Docker image mcr.microsoft.com/dotnet/aspnet:7.0-bullseye-slim
• Chromium 108

Other information

No response

Tagging subscribers to this area: @dotnet/area-system-diagnostics-process
See info in area-owners.md if you want to be subscribed.

Issue Details

---

Description

Hello. We have a cloud docker microservice which runs headless chromium in process. When we switch to .NET 7 we have encountered Process terminated. Error while reaping child. errno = 10 during killing chromium process. We cannot reproduce this behavior with .NET 6. I have prepared repro https://github.com/JaroslavMajera/DotNetReaper.

Thanks

Reproduction Steps

1. Prepare Dockerfile which periodically runs chromium or maybe something else which spawn child processes.
2. Start chromium process
3. Call Process.Kill a Process.WaitForExit()
4. Repeat

Source code: DotNetReaper-main.zip

Expected behavior

No Exception and chromium is successfully killed

Actual behavior

Sometimes we get

Process terminated. Error while reaping child. errno = 10 at System.Environment.FailFast(System.String) at System.Diagnostics.ProcessWaitState.TryReapChild(Boolean) at System.Diagnostics.ProcessWaitState.CheckChildren(Boolean, Boolean) at System.Diagnostics.Process.OnSigChild(Int32, Int32)

Regression?

Cannot reproduce on .NET 6

Known Workarounds

No response

Configuration

• .NET 7
• Docker image mcr.microsoft.com/dotnet/aspnet:7.0-bullseye-slim
• Chromium 108

Other information

No response

Author:	JaroslavMajera
Assignees:	-
Labels:	area-System.Diagnostics.Process
Milestone:	-

[adamsitnik]

@tmds is there any chance you could take a look?

[tmds]

I haven't run the reproducer yet, but I think it's a regression caused by #65061.

When the dotnet process runs as pid 1, it will reap also processes that are not direct children here:

runtime/src/libraries/System.Diagnostics.Process/src/System/Diagnostics/ProcessWaitState.Unix.cs

Lines 674 to 680 in e49d7ec

	if (reapAll)
	{
	do
	{
	pid = Interop.Sys.WaitPidExitedNoHang(-1, out _);
	} while (pid > 0);
	}

Inadvertently, we're also reaping direct children, which then leads to the FailFast.

If you change your ENTRYPOINT so dotnet is no longer pid 1, the issue should disappear. Something like:

ENTRYPOINT ["/bin/sh", "-c", "dotnet DotNetReaper.dll"]

I'll work on a fix next week.